Changing Parameters

The user can change model parameters by selecting Parameters from the Metrics menu.

The word-frequency file contain words in the target language with their associated frequency in a corpus. Entries in this file are used to calculate the keystrokes-per-character measurement and the words-per-minute performance prediction.

The words-per-minute prediction is based on a Fitts' Law model to calculate movement time between key. Slope and intercept values represent coefficients in Fitts' Law. The tmin value represents the time to repeat the last key press.

The presence of ambiguous words requires the user to expend perceptual and cognitive effort when using ambiguous keypads. The Model Human Processor (MHP) by Card, Moran and Newell models human mental processes and presents timing values representing the duration of mental operations. By incorporating such timing values into the equation for movement time, the WPM performance prediction is adjusted to account for this effort. The following is a description of each component:

tE
Time occupied with an eye movement
tP
Delay between the onset of a stimulus and perception of that stimulus
tM
Time to send a motor impulse
tC
Time to use the contents of short-term memory as input for an operation (e.g., accessing long-term memory) that change the contents of short-term memory
tS
Time to determine if the perceived word is the intended one
The user can select the preset values for an expert, typical, or novice user, or set custom values. To discount the effect of perceptual and cognitive effort (i.e., to use the traditional performance prediction equation) use custom values of zero for all components. For more information regarding the incorporation of MHP timing values into the WPM performance prediction, please see the TnToolkit's website.

Help Index